It follows from
Euler's formula that any simplicial 2-sphere with
n vertices has 3
n − 6 edges and 2
n − 4 faces. The case of
n = 4 is realized by the tetrahedron. By repeatedly performing the
barycentric subdivision, it is easy to construct a simplicial sphere for any
n ≥ 4. Moreover,
Ernst Steinitz gave a
characterization of 1-skeleta (or edge graphs) of convex polytopes in
R3 implying that any simplicial 2-sphere is a boundary of a convex polytope.
Branko Grünbaum constructed an example of a non-polytopal simplicial sphere (that is, a simplicial sphere that is not the boundary of a polytope).
Gil Kalai proved that, in fact, "most" simplicial spheres are non-polytopal. The smallest example is of dimension
d = 4 and has
f0 = 8 vertices. The
upper bound theorem gives upper bounds for the numbers
fi of
i-faces of any simplicial
d-sphere with
f0 =
n vertices. This conjecture was proved for simplicial convex polytopes by
Peter McMullen in 1970 and by
Richard Stanley for general simplicial spheres in 1975. The '''
g-conjecture'
, formulated by McMullen in 1970, asks for a complete characterization of f
-vectors of simplicial d
-spheres. In other words, what are the possible sequences of numbers of faces of each dimension for a simplicial d
-sphere? In the case of polytopal spheres, the answer is given by the '
g-theorem''', proved in 1979 by Billera and Lee (existence) and Stanley (necessity). It has been conjectured that the same conditions are necessary for general simplicial spheres. The conjecture was proved by
Karim Adiprasito in December 2018. == See also ==